The Interpretation of Population Change

Abstract

Various modifications and elaborations of the interpretation of the key factor methods of census data analysis are discussed. From the consideration of simple graphs it is shown that the length of the population cycle will influence the sign and value of regression coefficient in the Morris key factor analysis. Regular short cycles (oscillations) can arise as the result of a very sensitive density dependent mechanism. It is shown that with high rates of natality (e.g. 50 or 100 progeny/adult) such sensitive mechanisms could occur in nature without demanding excessively high survival fractions for population fluctuations of up to a hundred-fold. The effect of cycles of few generations is illustrated with a theoretical population with known density dependence, and with a natural population (the number of queen wasps reproducing). It is concluded that the pattern introduced by cycles of a few generations interferes with the Morris method and that the use of the method should be confined to populations that are increasing or decreasing over several generations. Attention is drawn to the fact that an age interval mortality may be both regulating (indicated by its regression on density) and disturbing (due to the variance around the regression line); over several generations the mean level of regulation by such a factor could theoretically be perfect, in spite of its short-term disturbing influence. A method is described for the separation of the roles of variations in natality and variations in mortality in determining the size of the resulting population. Using this technique, data from a variety of insects are analysed and it is shown that variations in natality dominate subsequent population size in some insects of agricultural habitats.